Transcellular Passage of Neisseria meningitidis across a Polarized Respiratory Epithelium

Bacterial strains and growth conditions. N. meningitidis was cultured in brain heart infusion (BHI) broth (Oxoid) or on BHI agar (1.5% [wt/vol], Agar No.1; Oxoid) with 5% Levinthal's supplement. Escherichia coli was grown in Luria-Bertani (LB) broth (Gibco) at 37°C with shaking or on LB agar. Details of the strains used in this work are shown in Table 1. Antibiotics were used at the following concentrations: kanamycin at 50 and 100 μg ml⁻¹ and erythromycin at 200 and 2 μg ml⁻¹ for E. coli and N. meningitidis, respectively, and carbenicillin at 50 μg ml⁻¹.

Cell culture.Calu-3 epithelial cells were cultured in DMEM:F12 (Invitrogen), while Chang epithelial cells were propagated in RPMI 1640 (Invitrogen). Media were supplemented with Glutamax (Invitrogen) and 10% fetal calf serum (PAA Laboratories). All cells were grown at 37°C in 5% CO₂. Calu-3 monolayers were grown on ∼1-cm², 1-μm-pore-size, BD Falcon cell culture inserts (BD Bioscience) containing polyethylene terephthalate membranes in 12-well tissue culture companion plates (BD Bioscience). Cells (4 × 10⁵) were seeded onto the apical side of membranes. For assays performed with liquid-covered culture (LCC), layers were maintained with 1 ml of culture medium in the apical chamber and 1.5 ml in the basolateral chamber. Cells were allowed to grow and differentiate for 5 days, with the media changed every second day. Air interface culture (AIC) monolayers were seeded and maintained as described above except that the apical media were removed after 2 days. The transepithelial electrical resistance (TEER) across monolayers was measured with an EVOM TEER meter attached to STX chopstick electrodes (World Precision Instruments); only monolayers with TEER of more than 1,400 Ω were used in experiments.

Texas Red-labeled 70,000-molecular-weight dextran (70,000-MW dextran; final concentration, 0.1 μM; Invitrogen) was added to the apical chamber of cell culture wells. Samples were taken from the basolateral chamber at various times afterwards, and the fluorescence analyzed with a spectrophotometer (Varian Cary Eclipse). For microscopy, coverslips were seeded with Calu-3 cells at a density of 4 × 10⁵ cells/well and Chang cells at a density of 5 × 10⁴ cells/well in 24-well plates. Cells were grown to semiconfluence overnight.

Challenge of cells with bacteria.All inocula consisted of 1 ml of a bacterial suspension per tissue culture well except for infection of AIC layers, for which a volume of 200 μl per well was used. Bacteria were harvested after overnight culture on solid media and resuspended in 400 μl of phosphate-buffered saline (PBS). The concentration of the bacterial suspension was adjusted in culture media to give the desired multiplicity of infection (MOI). For traversal assays, media in the basolateral chambers of monolayers were replaced immediately prior to challenge. For competitive infections (2), inocula consisting of a 1:1 ratio of mutant to wild-type bacteria with a final MOI of 40 were prepared and added to the apical chamber. The ratio of bacterial strains in the inoculum was determined by serial dilution and plating to media with or without antibiotics. After 7 h, cell culture inserts containing infected monolayers were washed then transferred to new wells containing 1.5 ml of culture media in the basolateral chamber and left for 1 h. The ratio of strains in the basolateral chamber was determined by plating to solid media. This procedure was repeated at 23 h postchallenge. The competitive index was calculated by dividing the ratio of mutant to wild-type in the output by the ratio in the input and was expressed logarithmically as the log competitive index (logCI).

Adhesion assays were performed by inoculating polarized monolayers with bacteria at an MOI of 40. After 3 h the monolayers were rinsed with Hanks buffered saline solution (Invitrogen) and then treated with 1% saponin to lyse cells and release all cell-associated bacteria. The numbers of cell-associated bacteria were enumerated by plating onto solid BHI media. The influence of purified CD46 on adhesion was assessed by adding the purified protein to bacteria prior to infection of cells. The inoculum was prepared at the required MOI, and 6.25 μg of purified CD46 (consisting of the extracellular domain alone) was added per 10⁷ bacteria. Bacteria and CD46 were incubated together at 37°C for 30 min, and then added to the cells. When required, different anti-CD46 monoclonal antibodies (MAbs) or an isotype control at a concentration of 2 μg/ml were added to the infection medium for 3 h. Details of antibodies used in the present study are given in Table 2.

For recovery of intracellular bacteria, Calu-3 monolayers were challenged with bacteria at an MOI of 40. At 8 or 24 h after inoculation, the monolayers were washed three times with PBS on their apical and basolateral surfaces and then incubated with 200 μg of gentamicin/ml for 30 min, again in both the apical and the basolateral chambers. Next, cell culture inserts were washed three times with PBS, and cells were lysed by a 5-min treatment with 1% saponin in PBS. The number of intracellular bacteria was established by serial dilution of lysates and plating to solid media.

Chemicals and inhibitors.Taxol and nocodazole (Sigma) were dissolved in dimethyl sulfoxide (DMSO) and used at a final concentration of 2 μM. Nocodazole was added to media on the apical surfaces of cells, which were then incubated at 4°C for 45 min and left at 37°C for 45 min prior to challenge. Taxol was added to the cells at the time of the challenge. The number of bacteria traversing monolayers was expressed as the proportion of the number of bacteria recovered from the basolateral chamber of wells containing no cells. Latrunculin A (Sigma) was used at a final concentration of 1 μM. In all cases, the final concentration of DMSO in culture wells never exceeded 0.1%.

Fixation, labeling, and staining.Antibodies were diluted in 10% horse serum in PBS. Samples for confocal microscopy were washed twice with PBS, fixed for 15 min at room temperature in 3% paraformaldehyde in PBS, and then washed three times with PBS. For labeling of intracellular and extracellular bacteria, the cells were incubated for 1 h with an α lipopolysaccharide (LPS) antibody (αL3,7,9), after which the cells were washed twice in PBS and incubated for 30 min with an α-mouse RRX-conjugated secondary antibody to label extracellular bacteria. Cells were permeabilized by treatment with 0.2% Triton X-100 (TX-100) in PBS, and the labeling process was repeated except that a Cy2-conjugated α-mouse secondary antibody was used to label both intracellular and extracellular bacteria. Host cell structures were labeled after permeabilization; actin was stained with fluorescently labeled phalloidin. After labeling, coverslips were washed twice in PBS and then double-distilled H₂O, mounted in ProLong Gold (Invitrogen), and analyzed by fluorescence microscopy (BX40; Olympus) or a confocal laser scanning microscopy (LSM510; Zeiss). Images were imported into Adobe Photoshop CS2 for processing.

For transmission electron microscopy (TEM), samples were washed three times on ice with ice-cold PBS and then immersed in EM-grade glutaraldehyde (Agar Scientific) diluted to 0.5% in 200 mM sodium cacodylate (Agar Scientific) at room temperature for 30 min. Next, the samples were washed three times in 200 mM sodium cacodylate buffer and postfixed in 1% osmium tetroxide-1.5% potassium ferrocyanide at room temperature for 1 h. Cells were then washed in water, incubated in 0.5% magnesium-uranyl acetate overnight at 4°C, washed again in water, dehydrated in ethanol, and embedded in Epon. Cross-sections were cut through fixed monolayers and placed on EM grids. Lead citrate was added for contrast, and the grids were then examined by using a Biacore 2000 transmission electron microscope.

Tfp and capsule expression.Monolayers were infected with MC58 at an MOI of 40 as described above. After 24 h, the bacteria were recovered from the apical and basal chambers. A total of 10 colonies from the apical and 10 from the basal chambers were grown overnight on solid media and then harvested and resuspended in 0.5 ml of PBS (Sigma, United Kingdom). For slot blot analysis of Tfp expression, samples were combined with an equal volume of 2× SDS-PAGE sample buffer, boiled for 10 min, and then transferred to Immobilon P polyvinylidene fluoride (PVDF) membranes (Millipore) by using a Minifold slot blot system (Whatman). The membranes were blocked in 5% milk in PBS for 1 h at room temperature with gentle shaking. Relevant antibodies were diluted in PBSTM (PBS, 0.05% Tween 20, 0.5% skimmed milk) and incubated with the PVDF membrane for 1 to 2 h, followed by three washes with PBSTM for 10 min each time. After incubation with the primary MAb (SM1 which recognizes class I pilin [70]), the PVDF membrane was washed with PBSTM three times for 10 min each time. A goat anti-murine immunoglobulin antibody conjugated to horseradish peroxidase (DakoCytomation) was used as the secondary antibody. Antibody binding was detected by using ECL Western blotting detection reagents (GE Healthcare).

Capsule expression was analyzed by enzyme-linked immunosorbent assay (ELISA). Suspensions of bacteria were prepared in PBS as described above and diluted to a concentration of 10⁹. Bacteria were heated at 56°C for 1 h and then resuspended in ELISA coating buffer (Sigma). Samples (100 μl) were added to the wells of 96-well plates and left overnight at 4°C. Wells with coating buffer alone or those coated with MC58ΔsiaD were included as controls. Next, the plates were washed with water and then incubated for 2 h with 2% bovine serum albumin in PBS. After blocking, plates were washed three times with PBS-0.1% Tween and incubated with α-serogroup B capsule MAb for 1 h at 37°C. Plates were washed three times in PBS-0.1% Tween and incubated with α murine immunoglobulin goat polyclonal antibody (pAb) conjugated to horseradish peroxidase (αM HRP) for a further hour at 37°C. After further washes, 100 μl of Sigmafast OPD substrate (Sigma) was added to plates for 6 min, and the reaction stopped with 100 μl of 0.3 M HCl. The absorbance was read at 492 nm using Ascent software.

Interactions between Calu-3 respiratory epithelial cells and N. meningitidis.Initially, we compared the interactions of the meningococcus with Calu-3 and Chang cells grown to semiconfluence and then challenged with MC58, a clinical serogroup B isolate of N. meningitidis, at an MOI of 100, which is in the range of MOIs used in other studies. This MOI represents a midpoint in a large range of MOIs used to study the interaction of N. meningitidis with Chang cells (4, 73). MC58 adhered to and invaded into Calu-3 and Chang cells at similar levels (the average numbers of adherent bacteria per cell were 17.4 and 20 for Chang and Calu-3 cells, respectively; the average numbers of internalized bacteria per cell were 2.18 and 2.21 for Chang and Calu-3 cells, respectively [see Fig. S1 in the supplemental material]). Furthermore, adhesion of N. meningitidis to Calu-3 cells was facilitated by Tfp (see Fig. S1 in the supplemental material), in keeping with results using other epithelial cell lines (47, 50, 69), and Calu-3 cells express CEACAMS (results not shown), which are receptors for Neisseria Opas (75).

Since N. meningitidis adheres to and enters into Calu-3 cells, we next grew the cells on permeable membranes in cell culture inserts so they formed a barrier between the apical and basolateral chambers (see Fig. S2 in the supplemental material). The development of tight junctions between cells was monitored by measuring the TEER across layers. The level of TEER across Calu-3 cells was influenced by a number of factors. Reducing the size of pores in membranes led to higher levels of TEER but pores of <1 μm impeded the passage of bacteria (not shown). Therefore, we selected membranes with 1-μm pores for all subsequent studies, which, with seeding cells at a density of 4 × 10⁵ per cm², gave consistent levels of TEER of ∼1,850 Ω after 5 days (Fig. 1A). The development of the epithelial cell barrier after the seeding of wells was also demonstrated by the permeability of the monolayer to fluorescently labeled 70,000-MW dextran added to the apical chamber. Increasing TEER over time coincided with decreasing permeability of the layer to the tracer (Fig. 1A), demonstrating that the cells formed an effective barrier within 5 days of seeding.

• Open in new tab
• Download powerpoint

FIG. 1.

Morphology of Calu-3 monolayers. (A) Development of the Calu-3 monolayer over time, showing increasing TEER and decreasing permeability to Texas Red-labeled 70,000-MW dextran. Fluorescent tracer was added to the apical chamber of monolayers at various times postseeding, and the levels of fluorescence in the basolateral chamber were measured after 6 h. Error bars show the standard deviation from at least three separate experiments. (B) For TEM analysis, cells were grown for 5 days on cell culture inserts. A lower-magnification view (scale bar, 4 μm) shows cells organized into a monolayer of columnar cells joined by tight junctions (TJ) and desmosomes (D). The cell culture membrane (CCM) can be seen on the basolateral surface of the cells. Apical-basal differentiation is demonstrated by the basal location of nuclei and the presence of microvilli (MV) on the apical surface. (C) Analysis of the cell boundaries shows the formation of extensive junctional structures including full junctional complexes (JC) composed of (from apical to basal surface) a tight junction (TJ), an adhesion belt (AB), and a desmosome (D). Scale bar, 1 μm. (D) Close apposition of cell membranes within a tight junction. Scale bar, 1 μm. The integrity of the layer and presence of tight junctions was further confirmed by using confocal microscopy to visualize the tight junction-associated proteins, ZO-1 (E) and occludin (F). Scale bars, 5 μm.

Furthermore, TEM confirmed that Calu-3 cells formed confluent monolayers with the morphological features of a polarized respiratory epithelium (Fig. 1B, C, and D) consisting of a continuous monolayer of columnar cells linked by junctional structures. There was apical-basal polarization, with nuclei located toward the basolateral side of cells, the development of tight junctions, and microvilli on the apical surface (Fig. 1B). Full junctional complexes (with tight junctions, adhesion belts, and desmosomes) were present between cells (Fig. 1C), and the membranes of adjacent cells were barely distinguishable (Fig. 1D). Immunofluorescence microscopy of the tight junction proteins zonula occludens-1 (ZO-1; Fig. 1E) and occludin (Fig. 1F) showed the characteristic chicken-wire pattern formed by their continuous presence at the margins of cells. At this time, while the tracer diffused freely into the basolateral chamber in the absence of cells (see Fig. S3A in the supplemental material), no tracer was detected in the basolateral chamber even after 24 h (see Fig. S3B in the supplemental material). In addition, we challenged the apical surface of the monolayer with noninvasive E. coli DH5α and monitored its appearance in the basolateral chamber over time. When no cells were present on the cell culture membrane, E. coli was detected in the basolateral chamber within 2 h of challenge (see Fig. S3C in the supplemental material). In contrast, no bacteria were detected in the presence of Calu-3 cell monolayers (see Fig. S3D in the supplemental material) even after 8 h.

N. meningitidis traverses the epithelial cell barrier by a transcellular route.Next, we investigated the traversal of N. meningitidis across Calu-3 monolayers 5 days after seeding. The apical surfaces of layers were challenged with MC58 at an MOI of 40 and, 8 and 24 h later, the inserts containing infected membranes were washed thoroughly in media and then transferred into new wells. After 1 h, samples were taken from the basal chambers of the new wells and plated to BHI. N. meningitidis was detected traversing in small numbers (equivalent to 0.0035% of input) at 8 h postchallenge. By 24 h, the number of bacteria traversing over the course of 1 h had increased by around 4 orders of magnitude (Fig. 2A). Bacteria could occasionally be found in the basal chamber prior to 8 h but never earlier than 6 h (not shown). At 16 h, intermediate levels of traversing bacteria were found (not shown).

• Open in new tab
• Download powerpoint

FIG. 2.

N. meningitidis traverses the Calu-3 monolayer without affecting barrier function. (A) Polarized monolayers were challenged with MC58 at an MOI of 40. The number of bacteria traversing the layer was established by moving inserts to a fresh well at 8 and 24 h and then sampling 1 h later, from the basolateral chamber of the new well. The results are shown as the percentage of bacteria in the input. Dots represent individual wells from four separate experiments; horizontal bars represent the mean. (B) TEER was monitored in the presence or absence of bacteria for 24 h following inoculation. Despite bacterial passage, TEER did not fall below 40% of initial levels. (C) Texas Red-labeled 70,000-MW dextran was added to the apical side of cell monolayers, which were then challenged with MC58 at an MOI of 40, and the fluorescence in the basolateral chamber was monitored over the next 24 h. The fluorescence did not increase despite bacterial passage. Diffusion of tracer across inserts with no cells was included as a control. Error bars show the standard deviation from at least three experiments.

Traversal of the monolayer by N. meningitidis might result from loss of integrity of the tight junctions or cell death. Therefore, the TEER across the monolayer was monitored during the course of an infection. Although the TEER across the layer fell after bacterial challenge, the levels never dropped to below 40% of initial readings and also fell in wells infected with E. coli and the uninfected control wells (Fig. 2B). The integrity of the layer during infection was also monitored by adding the fluorescent tracer to the apical surface of the monolayer at the time of challenge. Although bacteria successfully traversed the monolayer, there was no increase in fluorescence in the basolateral chamber beneath the infected cells even after 24 h. Indeed, there was no significant difference in the level of fluorescence in the basolateral chamber of infected and uninfected monolayers (Fig. 2C).

Infected monolayers were also fixed and examined for the presence of the tight junction protein ZO-1. The distribution of ZO-1 in infected cells was indistinguishable from uninfected monolayers (Fig. 3A). The protein was present at the margins of all cells, with no gaps or disruption detected. In addition, we examined infected monolayers by TEM. At 8 h after inoculation, microcolonies of N. meningitidis were present on the apical surface of the monolayer (Fig. 3B). These bacteria were tightly associated with the apical surface of the cells of the monolayers and were often found on pedestal-like protrusions of the apical membrane (Fig. 3B and C). After 24 h, the number of adherent bacteria had increased and covered a larger area of the apical surface of the monolayer (Fig. 3D). Meningococci were still closely adherent to the apical surface (Fig. 3E), and examination of the margins of cells revealed that the junctional structures were intact (Fig. 3F), a finding consistent with the results obtaining by ZO-1 labeling; even though gaps were evident between cells below the tight junctions, these were present in infected and uninfected monolayers (Fig. 1 and 3). Taken together, these data demonstrate that there is no loss of integrity of the Calu-3 monolayer coincident with traversal of N. meningitidis, indicating that the meningococcus does not cross the epithelial barrier by a paracellular route or after the loss of cells from the layer.

• Open in new tab
• Download powerpoint

FIG. 3.

Microscopic analysis of infected monolayers. (A) Monolayers were inoculated with MC58 expressing green fluorescent protein (GFP) and then fixed 24 h later. The tight-junction protein ZO-1 was labeled, and the layers were examined by confocal microscopy. Despite bacterial infection, the pattern of ZO-1 labeling was unaffected. Scale bars, 5 μm. (B and C) For TEM, monolayers were challenged with MC58 and fixed 8 or 24 h later. After 8 h the monolayer remained intact and appeared undamaged despite the presence of bacterial colonies on the apical surface. Bacteria adhered tightly to cells and were often on pedestallike protrusions from the apical membrane. (D) After 24 h, there was no apparent disruption to the integrity of the monolayer despite increased numbers of bacteria on the apical surface. (E) No gaps in the layer were observed. Adherent bacteria remained tightly associated with the apical membrane. (F) Enlargement of the region between cells revealed the presence of major junctional structures. The boxed regions in panel D indicate sections enlarged are in panels E and F.

These results suggest that N. meningitidis traverses the respiratory monolayer by a transcellular route, which predicts that the cells of the monolayer should harbor a population of intracellular bacteria during traversal. Therefore, monolayers of cells were challenged with MC58 and then, after 8 or 24 h, were exposed to gentamicin for 1 h to kill extracellular bacteria. Cells were lysed by treatment with saponin to release intracellular bacteria, which were recovered by plating lysates to solid media. A population of gentamicin-protected bacteria (representing ca. 0.1% of the input) was consistently present in cells at 8 h postchallenge, and this number increased by an order of magnitude by 24 h postchallenge (Fig. 4A). In contrast, no bacteria were recovered from gentamicin-treated membranes that had been infected with N. meningitidis in the absence of cells (not shown).

• Open in new tab
• Download powerpoint

FIG. 4.

Identification of a population of intracellular bacteria in infected Calu-3 monolayers. Monolayers of cells were challenged with N. meningitidis at an MOI of 40. (A) After 8 or 24 h, cells were treated with gentamicin to kill extracellular bacteria and then lysed with saponin to release intracellular bacteria, which were enumerated by plating of the lysates. Each spot represents the result from an individual well from three separate experiments. Horizontal bars represent the arithmetic mean. (B) Latrunculin A and trypsin were used to disrupt the monolayer before treatment with gentamicin. Treatment with these chemicals caused no significant reduction in the gentamicin-protected population of N. meningitidis. No bacteria were recovered after gentamicin treatment in the absence of Calu-3 cells. Error bars show the standard deviation from at least three separate experiments. (C) Representative images showing intracellular bacteria (indicated by pink arrowheads), present in the monolayer at 24 h postinoculation. Monolayers were infected with MC58 expressing GFP at an MOI of 40 and fixed after 24 h. Extracellular bacteria were labeled by using an αL3,7,9 LPS MAb, followed by a RRX-conjugated secondary antibody (red). Actin was stained with Alexa 647 phalloidin (blue). Scale bars, 5 μm. Boxed areas are enlarged in the lower panels of i and ii.

It was possible that the recovered bacteria survived gentamicin treatment by occupying extracellular sites between cells in the monolayer that were inaccessible to the antibiotic. Therefore, the infected monolayers were disrupted by treatment with trypsin for 10 min or latrunculin A for 20 min prior to incubation with gentamicin. Exposure of the monolayer to trypsin for 20 min is sufficient to remove all cells from the membrane, while treatment with latrunculin A for 15 min leads to a complete loss of TEER (not shown). Treatment with latrunculin or trypsin did not alter the number of CFU recovered from the layer after exposure to gentamicin (Fig. 4B), a finding consistent with the bacteria being in an intracellular compartment. In addition, we performed confocal microscopy of cross-sections of the monolayer at 24 h after bacterial challenge. This revealed the presence of N. meningitidis within cells, including bacteria located in the basal regions of cells (Fig. 4C).

Tfp and the polysaccharide capsule are required for traversal of the monolayer.For several pathogens, the mechanisms mediating adhesion to and entry into cells in monolayers are different for those described with semiconfluent cells (22, 45). Since Tfp have been implicated in meningococcal traversal of gastrointestinal epithelial cells (50), we examined the ability of a mutant lacking Tfp (MC58ΔpilE) to cross the respiratory epithelial cell barrier. The apical surface of Calu-3 monolayers were challenged with a 1:1 ratio of the wild-type strain and MC58ΔpilE, and the ratio of the strains exiting the layer was determined at 8 and 24 h postchallenge. At these times, inserts containing infected membranes were washed thoroughly in media and then transferred into new wells. Aliquots from the basolateral chambers of new wells were taken 1 h later and plated to media with or without antibiotics, and the competitive index (CI) of the mutant calculated. Wells with inserts lacking cells were challenged with the same inoculum as controls. We found that MC58ΔpilE had a significant defect for traversal across the Calu-3 monolayer after 24 h with a mean logCI of approximately −0.5 (Fig. 5A). To determine how Tfp contribute to traversal, we analyzed the adhesion of the pilus-deficient mutant and the wild-type strain to the polarized respiratory monolayer in separate wells. Consistent with the CI for MC58ΔpilE for traversal, the Tfp mutant had a significant defect for adhesion to the monolayer (P < 0.001, Fig. 5B). Furthermore, the adhesion defect of MC58ΔpilE was reflected in reduced recovery of intracellular bacteria from MC58ΔpilE-infected monolayers compared to MC58-infected cells (Fig. 5C).

• Open in new tab
• Download powerpoint

FIG. 5.

Tfp are important for traversal of the epithelial cell barrier. (A) Monolayers of Calu-3 cells were challenged with a 1:1 mixture of MC58 and MC58ΔpilE on the apical surface, and the ratio of the strains exiting the basolateral surface over 1 h was determined 8 and 24 h later. The mutant lacking Tfp had a significant defect for traversal. The results are shown as the log competitive index (logCI). Each spot represents the result from an individual well from three separate experiments. Horizontal bars represent the arithmetic mean. (B) Monolayers of Calu-3 cells were challenged with either MC58 or MC58ΔpilE. After 3 h the monolayers were washed and then lysed with 1% saponin to release all cell-associated bacteria. Enumeration of cell-associated bacteria by plating revealed a large reduction in the numbers of cell-associated bacteria for the pilus-deficient strain. (C) Intracellular bacteria were recovered by the lysis of cells with saponin after gentamicin treatment. MC58ΔpilE is present in reduced numbers within the cells of the monolayer after 8 and 24 h compared to MC58. No bacteria were recovered after gentamicin treatment in the absence of Calu-3 cells (data not shown). Significant differences are indicated: *, P < 0.05; **, P < 0.01; and ***, P < 0.001 (unpaired t test). (D) Slot blot analysis of pilin expression in 10 randomly selected colonies from the apical and basolateral chambers of wells after a 24-h infection with MC58. Eight of 10 colonies expressed pilus in the basolateral chamber compared to 10 of 10 colonies in the apical chamber.

It has been proposed that Tfp expression is downregulated after initial adhesion to host cells (12). Therefore, we examined pilus expression in bacteria following passage through the monolayer. Slot blot analysis using the MAb SM1 was performed on bacteria recovered from the apical and the basolateral chambers of wells challenged with MC58. While all 10 colonies from the apical chamber expressed Tfp, 8 of 10 colonies were positive for Tfp in the basolateral chamber (Fig. 5D). Identical findings were obtained when the experiment was performed on a separate occasion.

Recent work has indicated that the sialic acid polysaccharide capsule of N. meningitidis contributes to survival within epithelial cells (60). We hypothesized that the capsule is also involved in traversal. Therefore, we determined the CI for traversal of MC58ΔsiaD, which is unable to express a capsule due to loss of SiaD, the polysialic acid transferase (26). The unencapsulated strain had a marked defect in traversal (Fig. 6A); after 8 h the mean logCI of the MC58ΔsiaD mutant was −1.2, while by 24 h the logCI had fallen to almost −4. To understand the basis of this defect, we examined the adhesion of MC58ΔsiaD to monolayers. Consistent with previous work (26), we found that the capsule deficient strain had a significant increase in adhesion (approximately five fold) compared to MC58 (P < 0.001, Fig. 6B). We also recovered intracellular bacteria from MC58 and MC58ΔsiaD-infected monolayers 24 h after challenge. In contrast to adhesion, there was a significant reduction in the intracellular population of MC58ΔsiaD compared to MC58 at this later time point (P < 0.001, Fig. 6C). Furthermore, since capsule expression is downregulated during adhesion (12), we examined capsule expression in bacteria from the apical and basolateral chambers at 24 h postchallenge. All strains were found to be encapsulated by ELISA (Fig. 6D). Both the pilus- and the capsule-deficient strains grew in media and crossed membranes without cells as MC58 (not shown), so their attenuation cannot be ascribed to a growth defect or failure to pass through the pores in the membranes.

• Open in new tab
• Download powerpoint

FIG. 6.

The polysaccharide capsule is crucial to the traversal of N. meningitidis. (A) The apical surfaces of Calu-3 cells were infected with a 1:1 mixture of MC58 and MC58ΔsiaD, and the ratio of the strains exiting the basal surface over 1 h at 8 and 24 h was determined. The capsule-deficient strain had a significant defect for traversal. The data are shown as the log competitive index (logCI). Each spot represents the result from an individual well from three separate experiments. Horizontal bars represent the arithmetic mean. (B) Monolayers of Calu-3 cells were challenged with either MC58 or MC58ΔsiaD. After 3 h monolayers were washed and then lysed with 1% saponin to release cell-associated bacteria. Enumeration of cell-associated bacteria by plating revealed a large increase in the numbers of cell-associated MC58ΔsiaD compared to the wild-type strain. (C) Intracellular bacteria were recovered by the lysis of cells with saponin after gentamicin treatment. Despite its increased adhesion, MC58ΔsiaD is present in reduced numbers within the cells of the monolayer after 24 h compared to MC58. No bacteria were recovered after gentamicin treatment in the absence of Calu-3 cells (data not shown). Significant differences are indicated: *, P < 0.05; **, P < 0.01; and ***, P < 0.001 (unpaired t test). (D) Monolayers were infected with MC58 as described, and bacteria were recovered from the apical and basolateral chambers 24 h later. Individual colonies (10 from each chamber) were grown overnight on solid media, and analyzed for capsule expression by ELISA using an anti-serogroup B MAb at a variety of dilutions. The unencapsulated strain MC58ΔsiaD and the wild-type strain were included as controls. All strains sampled from apical and basolateral chambers were encapsulated. Spots represent the A₄₉₂ of each strain.

An intact microtubule network is required for successful traversal.Transcellular traversal of cells by bacteria should be dependent on host as well as bacterial factors. Since microtubules are required for long-range trafficking of vesicles, including those containing pathogens, such as Salmonella enterica serovar Typhimurium (53) and N. gonorrhoeae (76), we next examined whether disruption of the microtubule network using taxol or nocodazole affected meningococcal traversal. Taxol inhibits the depolymerization of tubulin subunits, freezing intact microtubules (30). In contrast nocodazole depolymerizes microtubules, leaving cells devoid of polymerized tubulin (11). Neither drug affected bacterial replication in vitro (see Fig. S4 in the supplemental material) or the permeability of the layer to fluorescent tracer, although some reduction in TEER was detected (see Fig. S5 in the supplemental material). To confirm the effect of these treatments on cells, monolayers were examined 24 h after drug treatment for the presence of microtubules and the tight junction protein occludin (see Fig. S6 in the supplemental material). Abnormalities in the microtubule networks of the cells of the monolayers, including cells undergoing abnormal cell division, were seen in taxol-treated cells (see Fig. S6 in the supplemental material). In contrast, β-tubulin labeling was diffuse throughout nocodazole treated cells, consistent with microtubule depolymerization (see Fig. S5 in the supplemental material). The distribution of occludin was unchanged after treatment with taxol or nocodazole, demonstrating (along with the TEER and tracer data) that the layer remained intact despite drug treatment. Treatment of cells with taxol or nocodazole led to a significant decrease in traversal of N. meningitidis at 8 and 24 h postchallenge (the P values for the taxol- and nocodazole-treated layers were <0.05 and <0.01, respectively [Fig. 7A]). Interestingly, treatment of the layers with taxol or nocodazole did not affect the number of intracellular bacteria at 8 or 24 h (Fig. 7B). Taken together, these data demonstrate that an intact microtubule network is necessary for successful traversal of the layer by N. meningitidis.

• Open in new tab
• Download powerpoint

FIG. 7.

Disruption of microtubules within cells of the monolayer reduces traversal of N. meningitidis. (A) Calu-3 monolayers were treated with nocodazole or taxol and then challenged with MC58 at an MOI of 40. The number of bacteria in the basolateral chamber was established by plating at 8 and 24 h postchallenge. Both nocodazole and taxol caused a significant reduction in the number of bacteria in the basolateral chamber at both time points compared to untreated layers. (B) Monolayers of Calu-3 cells were treated with nocodazole or taxol and then challenged. Intracellular bacteria were recovered by lysis of cells with 1% saponin after exposure to gentamicin for 30 min. There was no significant difference in the number of intracellular bacteria at 8 or 14 h for either drug treatment compared to untreated layers. No bacteria were recovered after gentamicin treatment in the absence of Calu-3 cells (data not shown). Error bars show the standard deviation from at least three separate experiments. Significant differences are indicated: *, P < 0.05; **, P < 0.01; and ***, P < 0.001 (unpaired t test).

Infection of Calu-3 monolayers grown under air interface culture (AIC).Calu-3 monolayers can be grown under an air interface culture (24). Cells were seeded as described above except the medium was removed from the apical chamber 2 days after seeding. Monolayers were then left for a further 3 days. Consistent with previous work (24), the TEER across monolayers with AIC was approximately half that seen with monolayers grown under liquid-covered culture (LCC) (Fig. 8A). TEM examination revealed that the AIC monolayers closely resemble LCC monolayers (Fig. 8B). Competitive infections comparing the traversal of MC58 with pilus-deficient or capsule-deficient mutants with MC58 were performed using the AIC monolayers. In contrast to LCC monolayers, no traversal had occurred at 8 h postinoculation. However, by 24 h some bacteria had traversed the monolayer. Both the pilus-deficient and the capsule-deficient mutants had significant defects for traversal (Fig. 8C and D). These results indicate that AIC monolayers could be adapted to investigate N. meningitidis infection and retain their selective barrier function during meningococcal traversal.

• Open in new tab
• Download powerpoint

FIG. 8.

Challenge of monolayers grown with an air interface (AI). (A) Monolayers were seeded as for liquid covered monolayers, but the medium was removed from the apical chamber after 2 days. After 5 days the levels of TEER in the AI monolayers were approximately half that seen for layers grown in medium. Significant differences are indicated: *, P < 0.05; **, P < 0.01; and ***, P < 0.001 (unpaired t test). (B) TEM of AIC monolayers at 5 days. The gross morphology of the layer is not significantly different from that seen in monolayers grown under standard conditions. A lower-magnification view (scale bar, 5 μm) shows cells organized into a monolayer of columnar cells joined by tight junctions (TJ) and desmosomes (D). The cell culture membrane (CCM) can be seen on the basolateral surface of the cells. Boxed sections are enlarged in panels i and ii (scale bar, 1 μm). (C) Competitive index of MC58ΔpilE across AI monolayers. In contrast to monolayers grown under medium, there was no traversal of N. meningitidis at 8 h. However, by 24 h some bacteria had traversed the layer. The pilus-deficient mutant has a defect for traversal compared to MC58. (D) CI of MC58ΔsiaD for traversal across AI monolayers. The capsule-deficient mutant had a large defect for traversal. The results are shown as the log competitive index (logCI). Each spot represents data from an individual well from three separate experiments. Horizontal bars represent the arithmetic mean.

Role of CD46 during adhesion of N. meningitidis to polarized respiratory epithelial monolayers.CD46 is the proposed receptor for meningococcal Tfp (33). Therefore, we examined whether CD46 is expressed by Calu-3 cells. Western blot analysis was performed on whole-cell lysates cells from Calu-3 cells, as well as HeLa cells, which is known to express CD46 (39). We found Calu-3 cells expressed CD46 at higher levels than HeLa or COS-7 cells (Fig. 9A). Next, we analyzed the distribution of CD46 in Calu-3 monolayers. CD46 was localized to the basolateral aspect of cells in the monolayer (Fig. 9B), an observation consistent with findings from other cell types and tissues (36, 41, 66). Furthermore, challenge of monolayers with N. meningitidis did not affect the basolateral distribution of CD46 (Fig. 9C), which was confined to the lateral and basolateral margin of cells.

• Open in new tab
• Download powerpoint

FIG. 9.

Role of CD46 during binding of N. meningitidis to polarized respiratory epithelial monolayers. (A) CD46 expression detected by Western blot analysis of lysates of cell lines (indicated). Full-length CD46 migrates with predicted molecular masses of between 50 and 79 kDa, while the purified SCR domains are detected as a band of 40 kDa. (B) Expression of CD46 in polarized monolayers examined by confocal microscopy. Five-day-old monolayers immunolabeled for CD46 detected with a fluorescein isothiocyanate (FITC)-conjugated, secondary antibody (green). Actin was counterstained with phalloidin conjugated to an Alexa 595 fluorophore (red). (C) Monolayers of Calu-3 cells were infected with MC58 expressing eGFP (green) at an MOI of 40 for 8 h, and immunolabeled CD46 was detected with a RRX (red)-conjugated secondary antibody; actin was stained using phalloidin conjugated to Alexa 647 (blue). Representative images are shown. Scale bars, 5 μm. (D) Monolayers of Calu-3 cells were challenged with MC58 at an MOI of 40 with or without 25 μg of purified CD46 per well (10⁵ cells). MC58 and purified CD46 were preincubated for 30 min prior to challenge. Monolayers were also infected with MC58 preincubated with DMEM:F12 as a control. (E) Calu-3 cells were infected with MC58 at an MOI of 40 in presence or absence of anti-human CD46 MAbs (indicated) at 2 μg/ml. Adhesion of N. meningitidis to cells was calculated by recovering bacteria 3 h later; error bars indicate the standard deviations of experiments performed on three occasions.

To test whether CD46 is involved in meningococcal adhesion to the monolayer, two different approaches were used. First, MC58 was incubated with purified CD46 (6.25 μg of CD46 per 10⁷ bacteria) for 30 min prior to infection of Calu-3 monolayers, and the adherent population was compared to bacteria incubated in the absence of the complement regulator. There was no significant difference in the level of adhesion of CD46-bound bacteria compared to bacteria that were not preincubated with CD46 (Fig. 9D). Furthermore, the rate of bacterial adhesion was determined with anti-CD46 MAbs in the infection medium (2 μg of anti-CD46 antibody per 1.6 × 10⁷ CFU). At 3 h after challenge, the adherent population of bacteria was compared to control wells without a MAb or with an isotype control. No significant difference in the adhesion of meningococci to the monolayer was observed in the presence of the MAbs (Fig. 9E). These data are consistent with CD46 playing little role in the adhesion of MC58 to the Calu-3 monolayers.